TURNER, HAENKE AND HUBBARD GLACIERS 109 and no conditions of vegetation known to us are opposed to this interpretation. The rapid retreat of these glaciers to their present positions is not impossible, judging by the rapid retreat of Nunatak Glacier from 1895 to 1909 and the recession of Muir and Grand Pacific Glaciers in Glacier Bay a mile and a half a year1 from 1794 to 1912. The absence of shorelines in northern Russell Fiord might be accounted for by the rapidity of retreat and quick shifting of lake levels, or by retreat southward instead of northward as suggested in a later paragraph. In 1888 Topham reports2 that "at the northeast end of Yakutat Bay the sea-water flows inland through a narrow passage frequently blocked by ice into a lake known as Disenchantment Bay. This bay is said to be 30 miles long and to be surrounded by high Tnonnt.fl.inH and great glaciers." In 1890 Russell went far enough north of Haenke Island in the Conoin to see beyond the bend into Russell Fiord which he explored to the head in 1891, finding Hubbard Glacier not far from its present position.8 The authors are still of the opinion that Russell, Gilbert, and Davidson misinterpreted Malaspina's map and description. We believe, however, that Hubbard Glacier may have extended some distance farther south in 1792 than now, ending nevertheless north of Haenke Island then and in 1794, 1807, and 1823. We are inclined to believe that the Russell Fiord Lake existed as recently as 1823, as the map shows specifically, although this implies that an east branch of Hubbard-Variegated Glacier, joined by the Nunatak and Hidden Glaciers, then extended over twenty miles southeast and south to within six miles of the present head of Russell Fiord. The evidence is conclusive that this condition actually existed in a very recent period, and it may have been as late as 1823. The Disenchantment Bay bifurcation of such an expanded ice tongue would be farther back in 1823 than the Russell Fiord branch because the latter ended temporarily in fresh water rather than salt, but soon melted back and was dismembered because salt water from the Disenchantment Bay side got at it. A Yakutat native told us in 1905 that his father-in-law, then still living, remembered when Nunatak Glacier extended to a point near Marble Point, about five miles southeast of Hubbard Glacier in Russell Fiord. This ail accords well with the idea of a great advance of all these glaciers in the late eighteenth or early nineteenth century, the Hubbard and Nunatak Glaciers having now melted further into the moutains than the adjacent Yakutat and Malaspina, because outwash deposits protect large parts of the last two from salt water and rapid melting. Changes Between 1891 and 1905. From several points of view the Hubbard Glacier was photographed by Russell in 1891 and by Gilbert in 1899, and both of these observers have described the glacier at the time of their observations, while Gilbert has made a study of the conditions in 1899 as compared with those in 1891. Photographs were also made by the Canadian Boundary surveyors in 1895, by Bryant in 1897, and by the U. S. Fish Commission in 1901. In 1905, and again in 1906, we occupied the exact stations from which some of Russell's and Gilbert's photographs were taken, and the comparison of conditions was described in our earlier reports.4 Recent retreat 1 Martin, Lawrence, Glaciers and International Boundaries, Scientific American Supplement, Vol. LXXVI. 1913, pp. 129, 136-138. »Topham, H. W., Proc. Roy. Geog. Soc., Vol. XX,, 1889, p. 425, i Russell, I. C., Nat. Geog. Mag., Vol. m, 1891, pp. 99-100 and PI. 9; ISth Ann. Kept., TL S. Geol. Survey, 1894, pp. 85, 89. « Tarr, R. S. and Martin, Lawrence, Glaciers and Glaciation of Yakutat Bay, Alaska, Bull. Amer. Geog. Soc., Vol. XXXVm, 1906, pp. 154-155; Tarr, R. S., Professional Paper 64, U. S. Geol. Survey, 1909, p. 45.